Prior to conducing the survey, we adapted the 46-item “Access Barriers Checklist: Advocates” instrument, a cross-disability measure developed by the Oregon Institute on Development and Disability based on their systematic literature review (Rehabilitation Research Center on Health and Wellness for Persons with Long-term Disabilities, 2008 ). The items—yes/no checklists—were organized under the categories of transportation; availability and access of service or system; insurance; access and accommodation within facilities; social, family, and caregiver support; and individual (e.g. finding the medical system too confusing or fear). We felt many of the items had good face validity, but we were concerned that the language might not be accessible to some autistic individuals and that the items did not address some important autism-related barriers. As a group, the AASPIRE team adapted the items to make them more accessible to autistic participants by clarifying language or sentence structure and adding pop-up definitions for difficult words. We also removed redundant items, added items we felt were important autism-specific barriers—such as sensory discomforts, difficulty identifying symptoms, or concerns about melt-downs—and replaced the single item on communication with a new, more detailed, section of communication-related barriers. We also added “other” options in each section for barriers not listed. The final instrument included in the healthcare survey had 60 potential barriers, as well as six options for “other” (one per category: transportation and access to services; insurance; access and accommodation within facilities; social, family, and caregiver support; individual level; communication).
After data collection, we reduced the pool of 66 barrier items on our healthcare survey to the 41 barriers that were endorsed by 10% or more of participants in any of the three groups (i.e. those items that presented barriers to a non-trivial proportion of participants). To aid with clarity of data analysis and presentation, we qualitatively sorted the 41 barriers into semantically related categories: (1) emotional, (2) executive function, (3) healthcare navigation, (4) provider attitudes, (5) patient-provider communication, (6) sensory, (7) socio-economic, (8) support, and (9) waiting. We consider these 41 items to be the “Barriers to Healthcare Checklist–Long Form” and present results for these items here.
We also wished to create a Short Form version that would be more practical to use in clinical or research settings. We combined functionally redundant items; for example, “Transportation costs too much,” and “I live in rural areas or the doctor’s office is too far away” became, “I do not have a way to get to my doctor’s office.” We also combined items at a low level of granularity into a higher level of granularity or dropped lower granularity items in favor of higher granularity ones; for example, separate items about sensory issues in facilities, sensory issues affecting communication, and sensory issues impacting tests and exams became, “Sensory discomforts (for example, the lights, smells, or sounds) get in the way of my healthcare.” We collapsed some related items into single items; for example, separate items about “fear and anxiety,” “embarrassment,” and “frustration or anger” became a single item “Fear, anxiety, embarrassment, or frustration keeps me from getting primary care.” The final items—the Barriers to Healthcare Checklist–Short Form—are shown with the original items they were derived from in
To verify our qualitative item reduction and Short Form creation process, we calculated pairwise correlations between those items we had grouped together to confirm that they were highly correlated with each other (or uncorrelated, for items for which a correlation would be counter-intuitive such as having no insurance and insurance not covering certain services).
